design of high-strength bainitic...
TRANSCRIPT
1
Design of novel high-strength bainitic steels. Part II
F. G. Caballero, H. K. D. H. Bhadeshia, K. J. A. Mawella, D. G. Jones and P. Brown
Dr F. G. Caballero and Professor H. K. D. H. Bhadeshia are in the Department of
Materials Science and Metallurgy, University of Cambridge, Pembroke Street,
Cambridge CB2 3QZ, UK. Dr K. J. A. Mawella, D. G. Jones and Dr P. Brown are in
the Structural Materials Centre, Defence Evaluation Research Agency, R1079,
Bldg A7, Farnborough GU14 0LX, UK
2
A combination of thermodynamic, kinetic and mechanical property models and
physical metallurgy principles was used in Part I of this study, to propose a number
of alloys which exploit the carbide-free bainitic microstructure at its theoretical best.
These alloys have been manufactured and the present paper (Part II) reports the
results of metallographic characterisation and mechanical tests. The proposed steels
are found to have the highest ever combination of strength and toughness for bainitic
microstructures, matching even the maraging steels which are at least thirty times
more expensive. The work confirms the alloy design procedures explained in Part I.
3
1 Introduction
A number of carbide-free bainitic steels were proposed for manufacturing in Part I of
this study. The steels were designed in the basis that the fraction of bainitic ferrite
obtained by continuous cooling transformation should be high enough to avoid any
large and unstable regions of high-carbon retained austenite, having only the films of
retained austenite to separate the bainite platelets. The steels also had to meet certain
hardenability and strength requirements.
The work presented here (Part II) is concerned with the microstructural and
mechanical property characterisation of the three bainitic steels proposed in Part I of
this study. As will be seen later, the results are exciting.
2 Experimental procedure
The actual chemical compositions of the steels studied are given in Table 1. Alloys
were prepared as 35 kg vacuum induction melts using high purity base materials.
After casting and cropping, the ingots were hot forged down to a thickness of 65 mm.
These were then homogenised at 1200 oC for 2 days and cut into smaller 65 mm
thick samples which were then forged down to 50 mm thickness. These samples
were then held at 900 oC for 2 hours, removed from the furnace and immediately hot-
pressed to a thickness of 25 mm before their temperature fell below 750 oC. Finally
they were allowed to cool in air. The particular deformation route used here (Fig. 1)
is consistent with a proprietary manufacturing standard.
4
2.1 MICROSTRUCTURAL EXAMINATION OF STEELS
Quantitative X-ray analysis was used to determined the fraction of retained austenite.
For this purpose, samples were cut from undeformed regions of Charpy specimens.
After grinding and final polishing using 0.25 µm diamond paste, the samples were
etched to obtain an undeformed surface. They were then step-scanned in a Philips -
PW 1730 X-ray diffractometer using unfiltered Cu Kα radiation. The scanning speed
(2θ) was 1 degree min-1. The machine was operated at 40 kV and 40 mA. The
retained austenite content was calculated from the integrated intensities of (200),
(220) and (311) austenite peaks, and those of (002), (112) and (022) planes of
ferrite.1 Using three peaks from each phase avoids biasing the results due to any
crystallographic texture in the samples.2 The carbon concentration in the austenite
was estimated by using the lattice parameters of the retained austenite.3
Specimens for transmission electron microscopy (TEM) were machined down to
3 mm diameter rod. The rods were sliced into 100 µm thick discs and subsequently
ground down to foils of 50 µm thickness on wet 800 grit silicon carbide paper. These
foils were finally electropolished at room temperature until perforation occurred,
using a twin-jet electropolisher set at a voltage of 40 V. The electrolyte consisted of
5 % perchloric acid, 15 % glycerol and 80 % methanol. The foils were examined in a
JEOL JEM-200 CX transmission electron microscope at an operating voltage of
200 kV.
Optical and scanning electron microscopy (SEM) were used to examine the etched
microstructures. Specimens were polished in the usual way and etched in 2% Nital
solution, and examined using a JEOL JXA-820 scanning electron microscope
operated at 10-15 kV. The volume fraction of bainite (Vb) was estimated by a
5
systematic manual point-counting procedure on scanning electron micrographs.4 A
grid superimposed on the microstructure provides, after a suitable number of
placements, an unbiased statistical estimate of Vb.
2.2 MECHANICAL PROPERTIES DETERMINATION
Tensile testing was carried out in accordance with BS EN 10 002-1: 1990 at room
temperature on a 100 kN Instron-6025 machine at a crosshead speed of 2 mm min-1.
Two specimens were tested for each alloy.
Impact toughness was measured at temperatures between 20 and -120 oC using a
300 J Charpy testing machine. Specimens were tested in accordance with BS EN 10
045-1: 1990. Six specimens were tested at each temperature for every alloy.
Compact tension specimens were used to measure values of plane strain fracture
toughness (KIc) at room temperature for the Ni1 and Ni2 alloys in accordance with
BS 7448 Part1: 1991. Compact tension / fracture toughness testing was done on a
100 kN ESH servohydraulic machine. The crosshead speed used was 1 mm min-1.
Fractography was carried on the Charpy impact toughness specimens using a JEOL
JXA-820 scanning electron microscope operating at 20 kV.
2.3 TEMPERED MICROSTRUCTURES
Austenite formation begins during heating at the Ac1 temperature and is completed
when the Ac3 temperature is reached. Tempering must be carried out below the Ac1
temperature to avoid the accidental formation of austenite. The austenite formation
temperatures were therefore determined using a Thermecmastor-Z thermomechanical
6
simulator. Cylindrical specimens 12 mm in height and 8 mm in diameter were heated
at a rate of 10 oC s-1 to 1000 oC and then cooled at 10 oC s-1. The Formation of
austenite during heating was detected by monitoring the fractional change in
dilatation with temperature. Table 2 shows the Ac1 and Ac3 temperatures. Note that
the Ac1 temperature is ill-defined in the present context because the microstructures
already contain some retained austenite. Ac1 should therefore be interpreted to mean
the temperature at which austenite growth begins during heating at the specified rate.
Specimens of the three alloys were tempered at temperatures ranging from 400 to
700 oC for an hour. Specimens for TEM were produced as described above.
3 Results and discussion
3.1 CHARACTERISATION OF MICROSTRUCTURE
Experimental data of the designed steel microstructures are presented in Table 3.
These data and the micrographs presented in Fig. 2 reveal that Ni1 and Ni2 have the
desired microstructure consisting of mainly bainitic ferrite and retained austenite.
Because of the high carbon content in austenite (xγ) for these alloys (Table 3), the
majority of residual austenite present after bainite is formed, is retained on cooling to
room temperature. In the SEM micrographs martensite regions are relatively
unetched, and appear light in colour. The thin interlath films of austenite, however,
have been etched away. It is difficult in thin-foil TEM experiments to judge whether
any martensite has formed as a result of the thinning process or whether it was
present in the original microstructure. To cope with this, the samples were tempered
7
at 200 oC for 2 h before making foils. Thus, martensite present in the original sample
would contain carbides, whereas any untempered martensite can be interpreted to be
austenite which underwent transformation during the foil preparation process. The
observation of tempered microstructures (Fig. 2.c) showed that the unetched regions
in Fig 2.b are martensitic. Note that the low temperature tempering (200 oC) heat
treatment does not affect bainitic ferrite since the latter does not contain excess
carbon.
Due to the high volume fraction of bainitic ferrite in Ni1 and Ni2 alloys, 0.62 and
0.81 respectively, the retained austenite was largely present as films between the
subunits of bainitic ferrite. Figure 3 shows typical bright-field images of carbide-free
upper bainite in Ni1 and Ni2 with interlath retained austenite films. These films have
a typical wavy morphology characteristic of the bainite in high-silicon steels
(Fig. 3.b).5-8
By contrast, the small amount of bainite in the Mn alloy (Table 3) causes much of the
residual austenite to transform to martensite during cooling because of lower carbon-
enrichment of the austenite (0.55 wt-%, Table 3). There is only 10 % of the residual
austenite retained to room temperature (Vγ=0.07, Table 3). These results are
consistent with the data on instability of residual austenite as function of austenite
carbon content illustrated elsewhere.9
The results of hardness tests are presented in Table 4. There is no doubt that the
microstructures of Ni1 and Ni2 alloys in the as-received condition consist mainly of
bainite and retained austenite. Their hardness values are closer to those of
microstructures obtained isothermally at 375 oC than the respective hardness values
of the martensite as determined by water quenched following austenitisation at
1000 oC for 15 min. Any difference between the hardness values of as-received and
8
isothermally transformed microstructures in Ni1 and Ni2 alloys is due to the different
degree of transformation to bainitic ferrite. Not surprisingly, the hardness value for
the as-received microstructure of Mn alloy (Fig. 2.a) is similar to that of a fully
martensitic sample.
3.2 MECHANICAL PROPERTIES
3.2.1 TENSILE STRENGTH AND DUCTILITY
Tensile test results are presented in Table 5. Plates of bainitic ferrite are typically
10 µm in length and about 0.2 µm in thickness (Fig. 3). This gives a rather small
mean free path for dislocation glide. Thus, the main microstructural contribution to
the strength of bainite is from the extremely fine grain size of bainitic ferrite.10
It is difficult to separate the effect of retained austenite on strength in these steels
from other factors. Qualitatively, austenite can affect the strength in several ways.
Residual austenite can transform to martensite during cooling to room temperature,
thus increasing the strength as observed in Mn alloy (Tables 3 and 5). On the other
hand, retained austenite interlath films can increase the strength by transforming to
martensite during testing, similar to the behaviour of TRIP (transformation induced
plasticity) steels.
The low yield/ultimate tensile strength ratios (YS/UTS) in Table 5 are due to the
presence of austenite and the generally large dislocation density in the
microstructure.11 Consequently, retained austenite increases the strain-hardening rate
of the steel. Likewise, tensile elongation is controlled by the volume fraction of
retained austenite.12 Retained austenite is a ductile phase compared to the bainitic
9
ferrite and would be expected to enhance ductility as far as the austenite is
homogeneously distributed along plate boundaries (film austenite). However,
isolated pools of austenite (blocky austenite) would influence unfavourably on both
elongation and UTS. From Table 5, it is clear that the steels present a combination of
high strength and good ductility. Moreover, tensile properties are much higher than
those required for defence applications (Table 3 in Part I).
3.2.2 IMPACT TOUGHNESS
Charpy impact test results are listed in Table 6 for all the alloys. The levels required
for defence applications in impact toughness have been also well achieved (Table 3
in Part I). A considerable improvement in toughness is obtained when the volume
fraction of bainite increases in the microstructure. This improvement occurs despite
the fact that the strength of the microstructures involved remains almost unchanged
(Table 5). From the data in Table 3, it is evident that the volume fraction of bainite
Vb and thus the carbon content of the retained austenite xγ explains the improvement
in toughness observed in Ni1 and Ni2 alloys as compared with the Mn alloy. The
results are consistent with the enhancement of toughness expected when the amount
of blocky austenite and martensite are reduced and, in general, when the stability of
residual austenite is increased. Figure 4 shows the fracture surfaces of the Charpy
impact specimens tested at room temperature. The two Ni alloys show ductile
dimpled fracture surfaces whereas the Mn alloy, which contains a large amount of
martensite, has many of the features of quasi-cleavage with isolated patches of
ductile fracture.
10
3.2.3 FRACTURE TOUGHNESS
A compact tension specimen design was used with thickness B=23.1 mm, width
W=46.5 mm, and crack length, a=24.5 mm, to width ratio of a/W=0.5. With the
measured yield stress of 1100 MPa, this would give a plane strain, KIc, measurement
capacity of ~ 300 MPa m . In the event, none of the specimens failed at sufficiently
low load to satisfy the KIc validity requirement.
The specimens showed a fairly significant non-linearity before maximum load. This
disqualified the use of KQ as a KIc measurement. The values quoted in Table 5 are for
the stress intensity at maximum load, Kmax. All the tests failed by microvoid
coalescence with no sign of cleavage. Tearing was stable at maximum load under
displacement control. Taking account of the non-linearity in the trace up to
maximum load gives the J-Integral at maximum load, Jmax and the stress intensity
KJmax calculated from Jmax values listed in Table 5.
Although the obtained fracture toughness results can not be considered valid, the
results are promising in that the steels tested are so tough that larger samples would
be needed to measure KIc. Toughness values of nearly 130 MPa m have been
obtained for strength in the range of 1600-1700 MPa. The good fracture toughness
obtained in these alloys is attributed to the presence of thin films of thermally and
mechanically stable interlath retained austenite. The role of retained austenite is to
refine the effective fracture grain size and to a blunt propagating crack.13
Figure 5 shows properties of mixed microstructures of bainitic ferrite and austenite,
versus those of quenched and tempered (QT) low-alloy martensitic alloys and
maraging steels.7 Small points in the graph refer to previous work,5-7,9,13 whereas the
two large points correspond to experimental data shown in Table 5. The alloys
11
designed theoretically in this work and produced by commercial continuous cooling
present the highest strength/toughness combinations ever recorded in bainitic steels.
These alloys show better mechanical properties of the QT low-alloy martensitic
alloys and match the critical properties of maraging steels, which are at least thirty
times more expensive.
3.3 CHARACTERISATION OF TEMPERED MICROSTRUCTURES
Figure 6 shows a plot of hardness values as a function of tempering temperature for
the three studied alloys. Tempering the designed microstructures of Ni1 and Ni2
steels at temperatures lower than 550 oC for an hour did not result in any significant
loss of hardness. However, tempering at 450 oC for an hour the as-received
microstructure of Mn alloy led to a drop in the hardness value from 597 to 555 HV.
Since the Mn alloy contains predominantly martensite (Vα' = 0.67, Table 3)
tempering at 400 oC for an hour is expected to lead to the precipitation of any excess
carbon. This is confirmed by Fig. 7.a which shows discrete carbide particles
precipitated inside a martensite plate in the Mn alloy. The retained austenite was still
intact with no significant signs of recovery in the bainitic ferrite.
Figures 7.b and c show bright-field images of microstructures obtained by tempering
at 400 oC for an hour in Ni1 and Ni2 alloys. Comparing tempering microstructures
with those as-received conditions (Fig. 3), it is clear that the tempering at 400 oC did
not result in any recovery in the bainitic ferrite (Fig. 7.b) or in the diffusional
decomposition of high carbon retained austenite (Fig. 7.c and d).
12
4 Conclusions
It has been demonstrated experimentally that models based on phase transformation
theory can be applied successfully to the design of high strength and tough steels.
These alloys designed to ensure that the hardenability of the steel is sufficient for
industrial production have achieved the highest strength and toughness combinations
to date for bainitic steels, at a cost thirty times less than that of maraging steels.
Likewise, it has been found that these two new steels also show an important
resistant to tempering. Steels have been tempered at temperatures ranging from 400
to 700 oC for an hour and no significant changes in hardness were detected in the
microstructure at tempering temperatures lower than 550 oC.
5 Acknowledgments
This work was carried out as part of Technology Group 4 (Materials and Structures)
of the MoD Corporate Research Programme. The authors would like to thank to
Professor Alan Windle for the provision of laboratory facilities at the University of
Cambridge.
13
6 References
1. J. DURNIN and K. A. RIDAL: Journal of the Iron and Steel Institute, 1968, 206,
60.
2. M. J. DICKSON: J. Appl. Cryst., 1969, 2, 176-180.
3. D. J. DYSON and B. HOLMES: Journal of the Iron and Steel Institute, 1970,
208, 469.
4. G. F. VANDER VOORT: ‘Metallography. Principles and Practice’, 427; 1984,
New York, McGraw-Hill.
5. V. T. T. MIIHKINEN and D. V. EDMONDS: Mater. Sci. Technol., 1987, 3, 422-
431.
6. V. T. T. MIIHKINEN and D. V. EDMONDS: Mater. Sci. Technol., 1987, 3, 432-
440.
7. V. T. T. MIIHKINEN and D. V. EDMONDS: Mater. Sci. Technol., 1987, 3, 441-
449.
8. L. C. CHANG: Metall. Trans., 1999, 30A, 909-916.
9. H. K. D. H. BHADESHIA and D. V. EDMONDS: Metal Sci., 1983, 17, , 411-
419.
10. K. J. IRVINE, F. B. PICKERING, W. C. HESELWOOD and M. J. ATKINS: J.
iron Steel Inst., 1957, 195, 54-67.
11. A. P. COLDREN, R. L. CRYDERMAN and M. SEMCHYSHEN: ‘Steel
Strengthening Mechanisms’, 17; 1969, Ann Arbor, USA, Climax Molybdenum.
12. B. P. J. SANDVIK and H. P. NEVALAINEN: Met. Tech., 1981, 15, 213-220.
13. H. K. D. H. BHADESHIA and D. V. EDMONDS: Metal Sci., 1983, 17, 420-425.
14
Table 1 Actual Chemical Composition of Designed Alloys, wt-%
Alloy C Si Mn Ni Cr Mo V
Mn 0.32 1.45 1.97 <0.02 1.26 0.26 0.10
Ni1 0.31 1.51 <0.01 3.52 1.44 0.25 0.10
Ni2 0.30 1.51 <0.01 3.53 1.42 0.25 <0.005
15
Table 2 Ac1 and Ac3 Temperatures in oC
Alloy Ac1 Ac3
Mn 807 881
Ni1 759 818
Ni2 775 835
16
Table 3 Quantitative Data on Microstructure and Hardness*
Alloy Vb Vγ Vα' xγ, wt-% Hardness, HV30
Mn 0.26±0.01 0.07±0.01 0.67±0.02 0.55 597±2
Ni1 0.62±0.05 0.12±0.01 0.26±0.04 0.92 493±5
Ni2 0.81±0.06 0.11±0.01 0.08±0.05 1.03 536±6
Vb bainitic ferrite volume fraction; Vγ retained austenite volume fraction; Vα'
martensite volume fraction; xγ carbon content in austenite
*Hardness values are averaged over 10 tests
17
Table 4 Hardness Data of Different Microstructures of the Alloys Studied
Alloy Heat treatment Hardness, HV30*
Mn As received microstructure
WQ†
350 oC for 30 min, WQ
597±2
605±5
467±7
Ni1 As received microstructure
WQ
375 oC for 30 min, WQ
493±5
647±8
426±4
Ni2 As received microstructure
WQ
375 oC for 30 min, WQ
536±6
669±7
423±9
*Hardness values are averaged over 10 tests
†WQ water quench
18
Table 5 Tensile* and Fracture Toughness† Properties
Alloy YS
MPa
UTS
MPa
Elongation
%
RA
%
Kmax
MPa m
Jmax
MPa m
KJmax
MPa m
Mn 1167 1790 13 44
Ni1 1150 1725 14 55 125 0.114 160
Ni2 1100 1625 14 59 128 0.134 174
*YS yield strength; UTS ultimate tensile strength; RA reduction of area.
†Kmax stress intensity factor at maximum load; Jmax J-integral at maximum load;
KJmax stress intensity values calculated from Jmax values.
19
Table 6 Charpy Impact Test Results*
Alloy Test temperature, oC Impact energy, J
Mn 20 34±1
-40 31±1
Ni1 20 58±2
-40 46±1
-120 34±2
Ni2 20 50±3
-40 43±1
-120 25±3
*Each value is a mean of six tests
20
Figure 1 Manufacturing route
21
a Mn alloy as received microstructure; b Ni1 alloy as received microstructure;
c Ni1 alloy as tempering at 200 oC for 2 hrs. microstructure; d Ni2 alloy as
received microstructure
Figure 2 Scanning electron micrographs of microstructures in the designed
alloys
22
a Ni1 alloy; b Ni2 alloy
3 Typical bright field images of microstructures formed by bainitic ferrite and
films of retained austenite
23
a Mn alloy, room temperature; b Ni1 alloy, room temperature; c Ni2 alloy, room
temperature
4 Fracture surfaces of Charpy impact specimens tested at different
temperatures. Scanning electron micrographs
24
5 Properties of mixed microstructures of bainitic ferrite and austenite, versus
those of quenched and tempered (QT) low-alloy martensitic alloys and
maraging steels
25
a Mn alloy; b Ni1 alloy; c Ni2 alloy
6 Plot of hardness values as a function of tempering temperature
26
a Discrete carbide particles precipitate inside a martensite plate in Mn alloy; b
None sings of recovery occur in the bainitic microstructure of Ni1 alloy; c Bright
field and d dark field images reveal that retained austenite is still intact in Ni2
alloy
7 Transmission electron micrographs of microstructures obtained by
tempering at 400 oC for an hour